<beerman>
are there any thoughts about this patch for for prtsweep? its a lot of shellcheck and linting fixes and a new mode (very quiet) in attempt to not have a prompt to press enter to make prtsweep useable in an external script: https://dpaste.com/CNNANQYT5
<beerman>
certainly helps if you are motivated to work with perl
<farkuhar>
these rewrites were inspired by the TODO file in the prtutils src tree, which indicated a long-run plan of switching to Perl.
<beerman>
it looks neater, I wouldn't care to switch to that. I see you already included README.md (cheers) - if there is a "no questions asked ever"-mode i'd be fine with that
<farkuhar>
also, the Perl versions are more conservative than their bash counterparts. They're more in keeping with the CRUX mantra "tools that just do the work for you, not the thinking"
<beerman>
is it scriptable? I have recently integrated prtsweep in my script that handles updating ports and thats a feature I'd miss and integrated with my patch
<farkuhar>
the Perl rewrites should work just fine when called from an external script. There's no interactivity, their behaviour is only governed by command-line options.
<beerman>
i'll check it out, would be nice if jaeger and jue would approve to make a switch
<farkuhar>
yeah, I'd leave it up to jue whether to replace the bash versions with my Perl rewrites, when releasing the next prt-utils tarball.
<beerman>
WARN: ./.signature not found, invalid port directory.
<beerman>
not everybody uses signatures all the time, it would be nice if they were optional
<farkuhar>
beerman: firefox builds just fine without wasi stuff. I don't mind keeping them around, though. It's low maintenance, versus requiring everyone who wants sandboxing to set up wasi-libc++ in their personal overlay.
<beerman>
alright, I'll move around them then, thanks. I am not going through all of them as its just too much..
<farkuhar>
thanks for the link regarding wayland-only builds of Firefox. I'll look into it soon.
<beerman>
sure, i am always glad to see big players make a move
<beerman>
i also wanted to take a look at https://peps.python.org/pep-0668/ and maybe implement that as well - feels like a good call to guide users into using venvs if they are so keen to bypass the ports system
<beerman>
farkuhar are you up for taking shaderc as well? libplacebo and gst-plugins-bad depend on it. quite frankly, if gstreamer wasn't that fixed set i'd add that to the to-drop-list as well as I am trying to minimize the packages I overtake to leave the field for fresh helpers. We just need to sort through all the contestants applying..
<farkuhar>
beerman: sure, I could adopt shaderc. It would be a shame to drop gstreamer and all its plugins, so thanks for continuing to maintain them.
<beerman>
not gstreamer as a whole, just plugins-bad which I don't have much use for
<beerman>
but sure, you do what you gotta do.. I guess
<jaeger>
I don't think I even use prtsweep... maybe as part of the automated ISO build
groovy2shoes has quit [Quit: Leaving]
groovy2shoes has joined #crux-devel
<farkuhar>
jaeger: it's been a while since I investigated, but I think one of the implications of FS#1313 is that you can get prtsweep-like results (a clean ports tree) just by syncing your repos with git, rather than the default rsync.
<farkuhar>
stenur or rmull might know more about that, if they do any building in their ports trees. I just kept using rsync out of habit, only switching to git for the contrib repo.